1. Field of the Invention
The present invention relates to an image forming apparatus and an image forming method for forming a visible image on a recording medium.
2. Description of the Related Art
Generally, if an image is expressed digitally, the image is expressed with multiple tones at 256 levels, for example. Furthermore, in the case of color images, such images are expressed with multiple tones for each color component, such as RGB or CMYK. However, since ordinary printers cannot express gradations at 256 levels, conventionally, a method has been adopted with which such multiple levels are reduced to a smaller number of levels (for example, two levels or four levels) that such printers can express, using a method called half toning.
For example, Japanese Patent Laid-Open No. 09-74488 (Patent Document 1) discloses technology for expressing an image having 256 levels with two to four levels, using a multi-level error diffusion method. Specifically, in a case in which 256 level data are converted into four level data, by using a threshold table to establish correspondence between 256 level data and four level data, the density data of a pixel of interest is converted into four level data using the multi-level error diffusion method. Then, the error (residual error) after this conversion into four level data is held, and used for adding to the peripheral pixels of the pixel of interest.
However, a dot of a pixel with no dots existing in the periphery, that is, an isolated pixel may become unstable. To cope with this, in order to reduce isolated pixels as much as possible, a method has been adopted with which contiguous pixels are made into a block and handled, and pixels are integrated in that block.
On the other hand, in the case in which pixels are integrated in a block in this way, there is a problem in that overloading may occur since such processing is complicated. As a method for reducing the load applied in such processing, in Japanese Patent Laid-Open No. 2001-309188 (Patent Document 2), image data corresponding to n pixels (n is a natural number) contiguous with each other in an inputted image is averaged, and multi-level error diffusion processing is performed on the averaged image data. Also, technology is disclosed with which based on the image data obtained by performing multi-level error diffusion processing, n pieces of image data expressed with N levels (N is a natural number) depending on a pattern set for each color are generated, and image data corresponding to n pixels is generated.
However, with conventional technology, in the case of averaging an inputted image for each group of a plurality of pixels contiguous with each other, if the difference between the densities of the pixels of such a plurality is great, information on the original image may disappear through averaging, so that an output image may deteriorate.